
# BASE
# ------------------------------------------------------
import numpy as np
import pandas as pd
import os
import gc
import warnings

# PACF - ACF
# ------------------------------------------------------
import statsmodels.api as sm

# DATA VISUALIZATION
# ------------------------------------------------------
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px


# CONFIGURATIONS
# ------------------------------------------------------
pd.set_option('display.max_columns', None)
pd.options.display.float_format = '{:.2f}'.format
warnings.filterwarnings('ignore')


# Import
train = pd.read_csv("D:/2021.2/store-sales-time-series-forecasting/train.csv")
test = pd.read_csv("D:/2021.2/store-sales-time-series-forecasting/test.csv")
stores = pd.read_csv("D:/2021.2/store-sales-time-series-forecasting/stores.csv")
#sub = pd.read_csv("../input/store-sales-time-series-forecasting/sample_submission.csv")   
transactions = pd.read_csv("D:/2021.2/store-sales-time-series-forecasting/transactions.csv").sort_values(["store_nbr", "date"])


# Datetime
train["date"] = pd.to_datetime(train.date)
test["date"] = pd.to_datetime(test.date)
transactions["date"] = pd.to_datetime(transactions.date)

# Data types
train.onpromotion = train.onpromotion.astype("float16")
train.sales = train.sales.astype("float32")
stores.cluster = stores.cluster.astype("int8")

train.head()


transactions.head(10)


temp = pd.merge(train.groupby(["date", "store_nbr"]).sales.sum().reset_index(), transactions, how = "left")
print("Spearman Correlation between Total Sales and Transactions: {:,.4f}".format(temp.corr("spearman").sales.loc["transactions"]))
px.line(transactions.sort_values(["store_nbr", "date"]), x='date', y='transactions', color='store_nbr',title = "Transactions" )

Spearman Correlation between Total Sales and Transactions: 0.8175


a = transactions.copy()
a["year"] = a.date.dt.year
a["month"] = a.date.dt.month
px.box(a, x="year", y="transactions" , color = "month", title = "Transactions")


a = transactions.set_index("date").resample("M").transactions.mean().reset_index()
a["year"] = a.date.dt.year
px.line(a, x='date', y='transactions', color='year',title = "Monthly Average Transactions" )


px.scatter(temp, x = "transactions", y = "sales", trendline = "ols", trendline_color_override = "red")


a = transactions.copy()
a["year"] = a.date.dt.year
a["dayofweek"] = a.date.dt.dayofweek+1
a = a.groupby(["year", "dayofweek"]).transactions.mean().reset_index()
px.line(a, x="dayofweek", y="transactions" , color = "year", title = "Transactions")


# Import 
oil = pd.read_csv("D:/2021.2/store-sales-time-series-forecasting/oil.csv")
oil["date"] = pd.to_datetime(oil.date)
# Resample
oil = oil.set_index("date").dcoilwtico.resample("D").sum().reset_index()
# Interpolate
oil["dcoilwtico"] = np.where(oil["dcoilwtico"] == 0, np.nan, oil["dcoilwtico"])
oil["dcoilwtico_interpolated"] =oil.dcoilwtico.interpolate()
# Plot
p = oil.melt(id_vars=['date']+list(oil.keys()[5:]), var_name='Legend')
px.line(p.sort_values(["Legend", "date"], ascending = [False, True]), x='date', y='value', color='Legend',title = "Daily Oil Price" )


temp = pd.merge(temp, oil, how = "left")
print("Correlation with Daily Oil Prices")
print(temp.drop(["store_nbr", "dcoilwtico"], axis = 1).corr("spearman").dcoilwtico_interpolated.loc[["sales", "transactions"]], "\n")


fig, axes = plt.subplots(1, 2, figsize = (15,5))
temp.plot.scatter(x = "dcoilwtico_interpolated", y = "transactions", ax=axes[0])
temp.plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[1], color = "r")
axes[0].set_title('Daily oil price & Transactions', fontsize = 15)
axes[1].set_title('Daily Oil Price & Sales', fontsize = 15);

Correlation with Daily Oil Prices
sales          -0.30
transactions    0.04
Name: dcoilwtico_interpolated, dtype: float64


a = pd.merge(train.groupby(["date", "family"]).sales.sum().reset_index(), oil.drop("dcoilwtico", axis = 1), how = "left")
c = a.groupby("family").corr("spearman").reset_index()
c = c[c.level_1 == "dcoilwtico_interpolated"][["family", "sales"]].sort_values("sales")

fig, axes = plt.subplots(7, 5, figsize = (20,20))
for i, fam in enumerate(c.family):
    if i < 6:
        a[a.family == fam].plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[0, i-1])
        axes[0, i-1].set_title(fam+"\n Correlation:"+str(c[c.family == fam].sales.iloc[0])[:6], fontsize = 12)
        axes[0, i-1].axvline(x=70, color='r', linestyle='--')
    if i >= 6 and i<11:
        a[a.family == fam].plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[1, i-6])
        axes[1, i-6].set_title(fam+"\n Correlation:"+str(c[c.family == fam].sales.iloc[0])[:6], fontsize = 12)
        axes[1, i-6].axvline(x=70, color='r', linestyle='--')
    if i >= 11 and i<16:
        a[a.family == fam].plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[2, i-11])
        axes[2, i-11].set_title(fam+"\n Correlation:"+str(c[c.family == fam].sales.iloc[0])[:6], fontsize = 12)
        axes[2, i-11].axvline(x=70, color='r', linestyle='--')
    if i >= 16 and i<21:
        a[a.family == fam].plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[3, i-16])
        axes[3, i-16].set_title(fam+"\n Correlation:"+str(c[c.family == fam].sales.iloc[0])[:6], fontsize = 12)
        axes[3, i-16].axvline(x=70, color='r', linestyle='--')
    if i >= 21 and i<26:
        a[a.family == fam].plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[4, i-21])
        axes[4, i-21].set_title(fam+"\n Correlation:"+str(c[c.family == fam].sales.iloc[0])[:6], fontsize = 12)
        axes[4, i-21].axvline(x=70, color='r', linestyle='--')
    if i >= 26 and i < 31:
        a[a.family == fam].plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[5, i-26])
        axes[5, i-26].set_title(fam+"\n Correlation:"+str(c[c.family == fam].sales.iloc[0])[:6], fontsize = 12)
        axes[5, i-26].axvline(x=70, color='r', linestyle='--')
    if i >= 31 :
        a[a.family == fam].plot.scatter(x = "dcoilwtico_interpolated", y = "sales", ax=axes[6, i-31])
        axes[6, i-31].set_title(fam+"\n Correlation:"+str(c[c.family == fam].sales.iloc[0])[:6], fontsize = 12)
        axes[6, i-31].axvline(x=70, color='r', linestyle='--')
        
        
plt.tight_layout(pad=5)
plt.suptitle("Daily Oil Product & Total Family Sales \n", fontsize = 20);
plt.show()


a = train[["store_nbr", "sales"]]
a["ind"] = 1
a["ind"] = a.groupby("store_nbr").ind.cumsum().values
a = pd.pivot(a, index = "ind", columns = "store_nbr", values = "sales").corr()
mask = np.triu(a.corr())
plt.figure(figsize=(20, 20))
sns.heatmap(a,
        annot=True,
        fmt='.1f',
        cmap='coolwarm',
        square=True,
        mask=mask,
        linewidths=1,
        cbar=False)
plt.title("Correlations among stores",fontsize = 20)
plt.show()


a = train.set_index("date").groupby("store_nbr").resample("D").sales.sum().reset_index()
px.line(a, x = "date", y= "sales", color = "store_nbr", title = "Daily total sales of the stores")


print(train.shape)
train = train[~((train.store_nbr == 52) & (train.date < "2017-04-20"))]
train = train[~((train.store_nbr == 22) & (train.date < "2015-10-09"))]
train = train[~((train.store_nbr == 42) & (train.date < "2015-08-21"))]
train = train[~((train.store_nbr == 21) & (train.date < "2015-07-24"))]
train = train[~((train.store_nbr == 29) & (train.date < "2015-03-20"))]
train = train[~((train.store_nbr == 20) & (train.date < "2015-02-13"))]
train = train[~((train.store_nbr == 53) & (train.date < "2014-05-29"))]
train = train[~((train.store_nbr == 36) & (train.date < "2013-05-09"))]
train.shape

(3000888, 6)

(2780316, 6)


c = train.groupby(["store_nbr", "family"]).sales.sum().reset_index().sort_values(["family","store_nbr"])
c = c[c.sales == 0]
c


print(train.shape)
# Anti Join
outer_join = train.merge(c[c.sales == 0].drop("sales",axis = 1), how = 'outer', indicator = True)
train = outer_join[~(outer_join._merge == 'both')].drop('_merge', axis = 1)
del outer_join
gc.collect()
train.shape

(2780316, 6)

(2698648, 6)


zero_prediction = []
for i in range(0,len(c)):
    zero_prediction.append(
        pd.DataFrame({
            "date":pd.date_range("2017-08-16", "2017-08-31").tolist(),
            "store_nbr":c.store_nbr.iloc[i],
            "family":c.family.iloc[i],
            "sales":0
        })
    )
zero_prediction = pd.concat(zero_prediction)
del c
gc.collect()
zero_prediction


c = train.groupby(["family", "store_nbr"]).tail(60).groupby(["family", "store_nbr"]).sales.sum().reset_index()
c[c.sales == 0]


fig, ax = plt.subplots(1,5, figsize = (20,4))
train[(train.store_nbr == 10) & (train.family == "LAWN AND GARDEN")].set_index("date").sales.plot(ax = ax[0], title = "STORE 10 - LAWN AND GARDEN")
train[(train.store_nbr == 36) & (train.family == "LADIESWEAR")].set_index("date").sales.plot(ax = ax[1], title = "STORE 36 - LADIESWEAR")
train[(train.store_nbr == 6) & (train.family == "SCHOOL AND OFFICE SUPPLIES")].set_index("date").sales.plot(ax = ax[2], title = "STORE 6 - SCHOOL AND OFFICE SUPPLIES")
train[(train.store_nbr == 14) & (train.family == "BABY CARE")].set_index("date").sales.plot(ax = ax[3], title = "STORE 14 - BABY CARE")
train[(train.store_nbr == 53) & (train.family == "BOOKS")].set_index("date").sales.plot(ax = ax[4], title = "STORE 43 - BOOKS")
plt.show()


a = train.set_index("date").groupby("family").resample("D").sales.sum().reset_index()
px.line(a, x = "date", y= "sales", color = "family", title = "Daily total sales of the family")


a = train.groupby("family").sales.mean().sort_values(ascending = False).reset_index()
px.bar(a, y = "family", x="sales", color = "family", title = "Which product family preferred more?")


d = pd.merge(train, stores)
d["store_nbr"] = d["store_nbr"].astype("int8")
d["year"] = d.date.dt.year
px.line(d.groupby(["city", "year"]).sales.mean().reset_index(), x = "year", y = "sales", color = "city")


holidays = pd.read_csv("D:/2021.2/store-sales-time-series-forecasting/holidays_events.csv")
holidays["date"] = pd.to_datetime(holidays.date)

# holidays[holidays.type == "Holiday"]
# holidays[(holidays.type == "Holiday") & (holidays.transferred == True)]

# Transferred Holidays
tr1 = holidays[(holidays.type == "Holiday") & (holidays.transferred == True)].drop("transferred", axis = 1).reset_index(drop = True)
tr2 = holidays[(holidays.type == "Transfer")].drop("transferred", axis = 1).reset_index(drop = True)
tr = pd.concat([tr1,tr2], axis = 1)
tr = tr.iloc[:, [5,1,2,3,4]]

holidays = holidays[(holidays.transferred == False) & (holidays.type != "Transfer")].drop("transferred", axis = 1)
holidays = holidays.append(tr).reset_index(drop = True)


# Additional Holidays
holidays["description"] = holidays["description"].str.replace("-", "").str.replace("+", "").str.replace('\d+', '')
holidays["type"] = np.where(holidays["type"] == "Additional", "Holiday", holidays["type"])

# Bridge Holidays
holidays["description"] = holidays["description"].str.replace("Puente ", "")
holidays["type"] = np.where(holidays["type"] == "Bridge", "Holiday", holidays["type"])

 
# Work Day Holidays, that is meant to payback the Bridge.
work_day = holidays[holidays.type == "Work Day"]  
holidays = holidays[holidays.type != "Work Day"]  


# Split

# Events are national
events = holidays[holidays.type == "Event"].drop(["type", "locale", "locale_name"], axis = 1).rename({"description":"events"}, axis = 1)

holidays = holidays[holidays.type != "Event"].drop("type", axis = 1)
regional = holidays[holidays.locale == "Regional"].rename({"locale_name":"state", "description":"holiday_regional"}, axis = 1).drop("locale", axis = 1).drop_duplicates()
national = holidays[holidays.locale == "National"].rename({"description":"holiday_national"}, axis = 1).drop(["locale", "locale_name"], axis = 1).drop_duplicates()
local = holidays[holidays.locale == "Local"].rename({"description":"holiday_local", "locale_name":"city"}, axis = 1).drop("locale", axis = 1).drop_duplicates()



d = pd.merge(train.append(test), stores)
d["store_nbr"] = d["store_nbr"].astype("int8")


# National Holidays & Events
#d = pd.merge(d, events, how = "left")
d = pd.merge(d, national, how = "left")
# Regional
d = pd.merge(d, regional, how = "left", on = ["date", "state"])
# Local
d = pd.merge(d, local, how = "left", on = ["date", "city"])

# Work Day: It will be removed when real work day colum created
d = pd.merge(d,  work_day[["date", "type"]].rename({"type":"IsWorkDay"}, axis = 1),how = "left")

# EVENTS
events["events"] =np.where(events.events.str.contains("futbol"), "Futbol", events.events)

def one_hot_encoder(df, nan_as_category=True):
    original_columns = list(df.columns)
    categorical_columns = df.select_dtypes(["category", "object"]).columns.tolist()
    # categorical_columns = [col for col in df.columns if df[col].dtype == 'object']
    df = pd.get_dummies(df, columns=categorical_columns, dummy_na=nan_as_category)
    new_columns = [c for c in df.columns if c not in original_columns]
    df.columns = df.columns.str.replace(" ", "_")
    return df, df.columns.tolist()

events, events_cat = one_hot_encoder(events, nan_as_category=False)
events["events_Dia_de_la_Madre"] = np.where(events.date == "2016-05-08", 1,events["events_Dia_de_la_Madre"])
events = events.drop(239)

d = pd.merge(d, events, how = "left")
d[events_cat] = d[events_cat].fillna(0)

# New features
d["holiday_national_binary"] = np.where(d.holiday_national.notnull(), 1, 0)
d["holiday_local_binary"] = np.where(d.holiday_local.notnull(), 1, 0)
d["holiday_regional_binary"] = np.where(d.holiday_regional.notnull(), 1, 0)

# 
d["national_independence"] = np.where(d.holiday_national.isin(['Batalla de Pichincha',  'Independencia de Cuenca', 'Independencia de Guayaquil', 'Independencia de Guayaquil', 'Primer Grito de Independencia']), 1, 0)
d["local_cantonizacio"] = np.where(d.holiday_local.str.contains("Cantonizacio"), 1, 0)
d["local_fundacion"] = np.where(d.holiday_local.str.contains("Fundacion"), 1, 0)
d["local_independencia"] = np.where(d.holiday_local.str.contains("Independencia"), 1, 0)


holidays, holidays_cat = one_hot_encoder(d[["holiday_national","holiday_regional","holiday_local"]], nan_as_category=False)
d = pd.concat([d.drop(["holiday_national","holiday_regional","holiday_local"], axis = 1),holidays], axis = 1)

he_cols = d.columns[d.columns.str.startswith("events")].tolist() + d.columns[d.columns.str.startswith("holiday")].tolist() + d.columns[d.columns.str.startswith("national")].tolist()+ d.columns[d.columns.str.startswith("local")].tolist()
d[he_cols] = d[he_cols].astype("int8")

d[["family", "city", "state", "type"]] = d[["family", "city", "state", "type"]].astype("category")

del holidays, holidays_cat, work_day, local, regional, national, events, events_cat, tr, tr1, tr2, he_cols
gc.collect()

d.head(10)


def AB_Test(dataframe, group, target):
    
    # Packages
    from scipy.stats import shapiro
    import scipy.stats as stats
    
    # Split A/B
    groupA = dataframe[dataframe[group] == 1][target]
    groupB = dataframe[dataframe[group] == 0][target]
    
    # Assumption: Normality
    ntA = shapiro(groupA)[1] < 0.05
    ntB = shapiro(groupB)[1] < 0.05
    # H0: Distribution is Normal! - False
    # H1: Distribution is not Normal! - True
    
    if (ntA == False) & (ntB == False): # "H0: Normal Distribution"
        # Parametric Test
        # Assumption: Homogeneity of variances
        leveneTest = stats.levene(groupA, groupB)[1] < 0.05
        # H0: Homogeneity: False
        # H1: Heterogeneous: True
        
        if leveneTest == False:
            # Homogeneity
            ttest = stats.ttest_ind(groupA, groupB, equal_var=True)[1]
            # H0: M1 == M2 - False
            # H1: M1 != M2 - True
        else:
            # Heterogeneous
            ttest = stats.ttest_ind(groupA, groupB, equal_var=False)[1]
            # H0: M1 == M2 - False
            # H1: M1 != M2 - True
    else:
        # Non-Parametric Test
        ttest = stats.mannwhitneyu(groupA, groupB)[1] 
        # H0: M1 == M2 - False
        # H1: M1 != M2 - True
        
    # Result
    temp = pd.DataFrame({
        "AB Hypothesis":[ttest < 0.05], 
        "p-value":[ttest]
    })
    temp["Test Type"] = np.where((ntA == False) & (ntB == False), "Parametric", "Non-Parametric")
    temp["AB Hypothesis"] = np.where(temp["AB Hypothesis"] == False, "Fail to Reject H0", "Reject H0")
    temp["Comment"] = np.where(temp["AB Hypothesis"] == "Fail to Reject H0", "A/B groups are similar!", "A/B groups are not similar!")
    temp["Feature"] = group
    temp["GroupA_mean"] = groupA.mean()
    temp["GroupB_mean"] = groupB.mean()
    temp["GroupA_median"] = groupA.median()
    temp["GroupB_median"] = groupB.median()
    
    # Columns
    if (ntA == False) & (ntB == False):
        temp["Homogeneity"] = np.where(leveneTest == False, "Yes", "No")
        temp = temp[["Feature","Test Type", "Homogeneity","AB Hypothesis", "p-value", "Comment", "GroupA_mean", "GroupB_mean", "GroupA_median", "GroupB_median"]]
    else:
        temp = temp[["Feature","Test Type","AB Hypothesis", "p-value", "Comment", "GroupA_mean", "GroupB_mean", "GroupA_median", "GroupB_median"]]
    
    # Print Hypothesis
    # print("# A/B Testing Hypothesis")
    # print("H0: A == B")
    # print("H1: A != B", "\n")
    
    return temp
    
# Apply A/B Testing
he_cols = d.columns[d.columns.str.startswith("events")].tolist() + d.columns[d.columns.str.startswith("holiday")].tolist() + d.columns[d.columns.str.startswith("national")].tolist()+ d.columns[d.columns.str.startswith("local")].tolist()
ab = []
for i in he_cols:
    ab.append(AB_Test(dataframe=d[d.sales.notnull()], group = i, target = "sales"))
ab = pd.concat(ab)
ab


d.groupby(["family","events_Futbol"]).sales.mean()[:60]

family                   events_Futbol
AUTOMOTIVE               0                  6.59
                         1                  6.20
BABY CARE                0                  0.15
                         1                  0.09
BEAUTY                   0                  4.02
                         1                  2.89
BEVERAGES                0               2577.73
                         1               2243.45
BOOKS                    0                  0.15
                         1                  0.00
BREAD/BAKERY             0                500.48
                         1                451.53
CELEBRATION              0                  9.06
                         1                  6.43
CLEANING                 0               1157.94
                         1               1102.08
DAIRY                    0                765.46
                         1                759.26
DELI                     0                286.35
                         1                264.20
EGGS                     0                185.05
                         1                180.72
FROZEN FOODS             0                167.31
                         1                134.66
GROCERY I                0               4079.68
                         1               3695.02
GROCERY II               0                 23.30
                         1                 22.24
HARDWARE                 0                  1.23
                         1                  1.22
HOME AND KITCHEN I       0                 22.16
                         1                 13.94
HOME AND KITCHEN II      0                 18.14
                         1                  6.19
HOME APPLIANCES          0                  0.49
                         1                  0.44
HOME CARE                0                190.43
                         1                159.24
LADIESWEAR               0                  9.54
                         1                  9.70
LAWN AND GARDEN          0                  6.94
                         1                  5.05
LINGERIE                 0                  7.74
                         1                  8.83
LIQUOR,WINE,BEER         0                 91.76
                         1                115.06
MAGAZINES                0                  3.18
                         1                  0.49
MEATS                    0                368.91
                         1                376.27
PERSONAL CARE            0                292.03
                         1                274.46
PET SUPPLIES             0                  4.24
                         1                  2.72
PLAYERS AND ELECTRONICS  0                  6.69
                         1                  4.57
POULTRY                  0                377.97
                         1                424.07
PREPARED FOODS           0                104.43
                         1                106.65
Name: sales, dtype: float32


# Time Related Features
def create_date_features(df):
    df['month'] = df.date.dt.month.astype("int8")
    df['day_of_month'] = df.date.dt.day.astype("int8")
    df['day_of_year'] = df.date.dt.dayofyear.astype("int16")
    df['week_of_month'] = (df.date.apply(lambda d: (d.day-1) // 7 + 1)).astype("int8")
    df['week_of_year'] = (df.date.dt.weekofyear).astype("int8")
    df['day_of_week'] = (df.date.dt.dayofweek + 1).astype("int8")
    df['year'] = df.date.dt.year.astype("int32")
    df["is_wknd"] = (df.date.dt.weekday // 4).astype("int8")
    df["quarter"] = df.date.dt.quarter.astype("int8")
    df['is_month_start'] = df.date.dt.is_month_start.astype("int8")
    df['is_month_end'] = df.date.dt.is_month_end.astype("int8")
    df['is_quarter_start'] = df.date.dt.is_quarter_start.astype("int8")
    df['is_quarter_end'] = df.date.dt.is_quarter_end.astype("int8")
    df['is_year_start'] = df.date.dt.is_year_start.astype("int8")
    df['is_year_end'] = df.date.dt.is_year_end.astype("int8")
    # 0: Winter - 1: Spring - 2: Summer - 3: Fall
    df["season"] = np.where(df.month.isin([12,1,2]), 0, 1)
    df["season"] = np.where(df.month.isin([6,7,8]), 2, df["season"])
    df["season"] = pd.Series(np.where(df.month.isin([9, 10, 11]), 3, df["season"])).astype("int8")
    return df
d = create_date_features(d)




# Workday column
d["workday"] = np.where((d.holiday_national_binary == 1) | (d.holiday_local_binary==1) | (d.holiday_regional_binary==1) | (d['day_of_week'].isin([6,7])), 0, 1)
d["workday"] = pd.Series(np.where(d.IsWorkDay.notnull(), 1, d["workday"])).astype("int8")
d.drop("IsWorkDay", axis = 1, inplace = True)

# Wages in the public sector are paid every two weeks on the 15 th and on the last day of the month. 
# Supermarket sales could be affected by this.
d["wageday"] = pd.Series(np.where((d['is_month_end'] == 1) | (d["day_of_month"] == 15), 1, 0)).astype("int8")

d.head(15)


d[(d.month.isin([4,5]))].groupby(["year"]).sales.mean()

year
2013   249.38
2014   285.53
2015   334.68
2016   489.90
2017   503.50
Name: sales, dtype: float32


pd.pivot_table(d[(d.month.isin([3]))], index="year", columns="family", values="sales", aggfunc="mean")


pd.pivot_table(d[(d.month.isin([4,5]))], index="year", columns="family", values="sales", aggfunc="mean")


pd.pivot_table(d[(d.month.isin([6]))], index="year", columns="family", values="sales", aggfunc="mean")


#a = d[d["store_nbr"]==1].set_index("date")
a = d[(d.sales.notnull())].groupby(["date", "family"]).sales.mean().reset_index().set_index("date")
for num, i in enumerate(a.family.unique()):
    try:
        fig, ax = plt.subplots(1,2,figsize=(15,5))
        temp = a[(a.family == i)]#& (a.sales.notnull())
        sm.graphics.tsa.plot_acf(temp.sales, lags=365, ax=ax[0], title = "AUTOCORRELATION\n" + i)
        sm.graphics.tsa.plot_pacf(temp.sales, lags=365, ax=ax[1], title = "PARTIAL AUTOCORRELATION\n" + i)
    except:
        pass


a = d[d.year.isin([2016,2017])].groupby(["year", "day_of_year"]).sales.mean().reset_index()
px.line(a, x = "day_of_year", y = "sales", color = "year", title = "Average sales for 2016 and 2017")


a = train.sort_values(["store_nbr", "family", "date"])
for i in [20, 30, 45, 60, 90, 120, 365, 730]:
    a["SMA"+str(i)+"_sales_lag16"] = a.groupby(["store_nbr", "family"]).rolling(i).sales.mean().shift(16).values
    a["SMA"+str(i)+"_sales_lag30"] = a.groupby(["store_nbr", "family"]).rolling(i).sales.mean().shift(30).values
    a["SMA"+str(i)+"_sales_lag60"] = a.groupby(["store_nbr", "family"]).rolling(i).sales.mean().shift(60).values
print("Correlation")
a[["sales"]+a.columns[a.columns.str.startswith("SMA")].tolist()].corr()

Correlation


b = a[(a.store_nbr == 1)].set_index("date")
for i in b.family.unique():
    fig, ax = plt.subplots(2,4,figsize=(20,10))
    b[b.family == i][["sales", "SMA20_sales_lag16"]].plot(legend = True, ax = ax[0,0], linewidth = 4)
    b[b.family == i][["sales", "SMA30_sales_lag16"]].plot(legend = True, ax = ax[0,1], linewidth = 4)
    b[b.family == i][["sales", "SMA45_sales_lag16"]].plot(legend = True, ax = ax[0,2], linewidth = 4)
    b[b.family == i][["sales", "SMA60_sales_lag16"]].plot(legend = True, ax = ax[0,3], linewidth = 4)
    b[b.family == i][["sales", "SMA90_sales_lag16"]].plot(legend = True, ax = ax[1,0], linewidth = 4)
    b[b.family == i][["sales", "SMA120_sales_lag16"]].plot(legend = True, ax = ax[1,1], linewidth = 4)
    b[b.family == i][["sales", "SMA365_sales_lag16"]].plot(legend = True, ax = ax[1,2], linewidth = 4)
    b[b.family == i][["sales", "SMA730_sales_lag16"]].plot(legend = True, ax = ax[1,3], linewidth = 4)
    plt.suptitle("STORE 1 - "+i, fontsize = 15)
    plt.tight_layout(pad = 1.5)
    for j in range(0,4):
        ax[0,j].legend(fontsize="x-large")
        ax[1,j].legend(fontsize="x-large")
    plt.show()


def ewm_features(dataframe, alphas, lags):
    dataframe = dataframe.copy()
    for alpha in alphas:
        for lag in lags:
            dataframe['sales_ewm_alpha_' + str(alpha).replace(".", "") + "_lag_" + str(lag)] = \
                dataframe.groupby(["store_nbr", "family"])['sales']. \
                    transform(lambda x: x.shift(lag).ewm(alpha=alpha).mean())
    return dataframe

alphas = [0.95, 0.9, 0.8, 0.7, 0.5]
lags = [16, 30, 60, 90]

a = ewm_features(a, alphas, lags)


a[(a.store_nbr == 1) & (a.family == "GROCERY I")].set_index("date")[["sales", "sales_ewm_alpha_095_lag_16"]].plot(title = "STORE 1 - GROCERY I");

	date	store_nbr	transactions
1	2013-01-02	1	2111
47	2013-01-03	1	1833
93	2013-01-04	1	1863
139	2013-01-05	1	1509
185	2013-01-06	1	520
231	2013-01-07	1	1807
277	2013-01-08	1	1869
323	2013-01-09	1	1910
369	2013-01-10	1	1679
415	2013-01-11	1	1813

	id	date	store_nbr	family	sales	city	state	type	cluster	IsWorkDay	holiday_national_binary	local_cantonizacio	local_fundacion	local_independencia	holiday_national_Primer_dia_del_ano
0	0	2013-01-01	1	AUTOMOTIVE	0.00	Quito	Pichincha	D	13	NaN	1	1	1	1	1
1	1782	2013-01-02	1	AUTOMOTIVE	2.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0
2	3564	2013-01-03	1	AUTOMOTIVE	3.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0
3	5346	2013-01-04	1	AUTOMOTIVE	3.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0
4	7128	2013-01-05	1	AUTOMOTIVE	5.00	Quito	Pichincha	D	13	Work Day	0	1	1	1	0
5	8910	2013-01-06	1	AUTOMOTIVE	2.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0
6	10692	2013-01-07	1	AUTOMOTIVE	0.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0
7	12474	2013-01-08	1	AUTOMOTIVE	2.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0
8	14256	2013-01-09	1	AUTOMOTIVE	2.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0
9	16038	2013-01-10	1	AUTOMOTIVE	2.00	Quito	Pichincha	D	13	NaN	0	1	1	1	0

Feature	Test Type	AB Hypothesis	p-value	Comment	GroupA_mean	GroupB_mean	GroupA_median	GroupB_median
events_Black_Friday	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	393.70	396.72	26.58	19.00
events_Cyber_Monday	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	472.65	396.57	24.00	19.00
events_Dia_de_la_Madre	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	386.35	396.74	17.00	19.00
events_Futbol	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	360.30	397.01	13.00	19.00
events_Terremoto_Manabi	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	519.38	394.30	30.00	19.00
holiday_national_binary	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	471.40	393.15	20.00	19.00
holiday_local_binary	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	472.89	396.41	24.00	19.00
holiday_regional_binary	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	254.61	396.76	11.00	19.00
holiday_national_Batalla_de_Pichincha	Non-Parametric	Fail to Reject H0	0.62	A/B groups are similar!	434.22	396.60	21.00	19.00
holiday_national_Carnaval	Non-Parametric	Fail to Reject H0	0.19	A/B groups are similar!	373.36	396.86	19.00	19.00
holiday_national_Dia_de_Difuntos	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	471.98	396.49	26.00	19.00
holiday_national_Dia_de_la_Madre	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	495.64	396.42	30.00	19.00
holiday_national_Dia_del_Trabajo	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	535.81	396.30	23.00	19.00
holiday_national_Independencia_de_Cuenca	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	528.54	396.40	23.37	19.00
holiday_national_Independencia_de_Guayaquil	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	476.40	396.52	27.00	19.00
holiday_national_Navidad	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	607.37	394.18	36.49	19.00
holiday_national_Primer_Grito_de_Independencia	Non-Parametric	Reject H0	0.02	A/B groups are not similar!	407.43	396.68	22.00	19.00
holiday_national_Primer_dia_del_ano	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	347.77	397.00	0.00	19.00
holiday_national_Viernes_Santo	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	327.01	396.92	16.00	19.00
holiday_regional_Provincializacion_Santa_Elena	Non-Parametric	Fail to Reject H0	0.20	A/B groups are similar!	188.21	396.72	11.50	19.00
holiday_regional_Provincializacion_de_Cotopaxi	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	245.20	396.73	5.00	19.00
holiday_regional_Provincializacion_de_Imbabura	Non-Parametric	Reject H0	0.01	A/B groups are not similar!	182.07	396.72	9.00	19.00
holiday_regional_Provincializacion_de_Santo_Do...	Non-Parametric	Fail to Reject H0	0.60	A/B groups are similar!	326.72	396.72	19.50	19.00
holiday_local_Cantonizacion_de_Cayambe	Non-Parametric	Fail to Reject H0	0.11	A/B groups are similar!	459.53	396.71	24.50	19.00
holiday_local_Cantonizacion_de_El_Carmen	Non-Parametric	Fail to Reject H0	0.07	A/B groups are similar!	235.88	396.72	17.50	19.00
holiday_local_Cantonizacion_de_Guaranda	Non-Parametric	Reject H0	0.01	A/B groups are not similar!	183.58	396.72	8.00	19.00
holiday_local_Cantonizacion_de_Latacunga	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	188.05	396.73	12.00	19.00
holiday_local_Cantonizacion_de_Libertad	Non-Parametric	Fail to Reject H0	0.28	A/B groups are similar!	297.36	396.71	12.00	19.00
holiday_local_Cantonizacion_de_Quevedo	Non-Parametric	Fail to Reject H0	0.13	A/B groups are similar!	213.73	396.72	18.50	19.00
holiday_local_Cantonizacion_de_Riobamba	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	195.55	396.72	8.00	19.00
holiday_local_Cantonizacion_de_Salinas	Non-Parametric	Fail to Reject H0	0.80	A/B groups are similar!	285.65	396.72	19.50	19.00
holiday_local_Cantonizacion_del_Puyo	Non-Parametric	Fail to Reject H0	0.33	A/B groups are similar!	150.41	396.72	23.00	19.00
holiday_local_Fundacion_de_Ambato	Non-Parametric	Fail to Reject H0	0.87	A/B groups are similar!	336.04	396.72	21.60	19.00
holiday_local_Fundacion_de_Cuenca	Non-Parametric	Fail to Reject H0	0.92	A/B groups are similar!	333.86	396.72	19.00	19.00
holiday_local_Fundacion_de_Esmeraldas	Non-Parametric	Fail to Reject H0	0.23	A/B groups are similar!	333.56	396.71	10.00	19.00
holiday_local_Fundacion_de_Guayaquil	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	288.48	396.80	14.00	19.00
holiday_local_Fundacion_de_Ibarra	Non-Parametric	Fail to Reject H0	0.05	A/B groups are similar!	224.81	396.72	7.50	19.00
holiday_local_Fundacion_de_Loja	Non-Parametric	Fail to Reject H0	0.26	A/B groups are similar!	441.52	396.71	20.00	19.00
holiday_local_Fundacion_de_Machala	Non-Parametric	Fail to Reject H0	0.50	A/B groups are similar!	297.36	396.72	17.00	19.00
holiday_local_Fundacion_de_Manta	Non-Parametric	Fail to Reject H0	0.63	A/B groups are similar!	255.17	396.72	15.00	19.00
holiday_local_Fundacion_de_Quito	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	732.59	396.15	41.00	19.00
holiday_local_Fundacion_de_Riobamba	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	193.41	396.72	3.00	19.00
holiday_local_Fundacion_de_Santo_Domingo	Non-Parametric	Fail to Reject H0	0.42	A/B groups are similar!	314.92	396.72	21.00	19.00
holiday_local_Independencia_de_Ambato	Non-Parametric	Fail to Reject H0	0.13	A/B groups are similar!	370.56	396.71	29.50	19.00
holiday_local_Independencia_de_Guaranda	Non-Parametric	Fail to Reject H0	0.06	A/B groups are similar!	219.94	396.72	6.00	19.00
holiday_local_Independencia_de_Latacunga	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	178.59	396.73	11.00	19.00
national_independence	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	457.03	396.07	23.00	19.00
local_cantonizacio	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	396.33	506.69	19.00	26.00
local_fundacion	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	396.82	250.79	19.00	14.00
local_independencia	Non-Parametric	Reject H0	0.00	A/B groups are not similar!	396.38	486.04	19.00	25.00

family	AUTOMOTIVE	BABY CARE	BEAUTY	BEVERAGES	BOOKS	BREAD/BAKERY	CELEBRATION	CLEANING	DAIRY	DELI	EGGS	FROZEN FOODS	GROCERY I	GROCERY II	HARDWARE	HOME AND KITCHEN I	HOME AND KITCHEN II	HOME APPLIANCES	HOME CARE	LADIESWEAR	LAWN AND GARDEN	LINGERIE	LIQUOR,WINE,BEER	MAGAZINES	MEATS	PERSONAL CARE	PET SUPPLIES	PLAYERS AND ELECTRONICS	POULTRY	PREPARED FOODS	PRODUCE	SCHOOL AND OFFICE SUPPLIES	SEAFOOD
year
2013	5.35	0.00	2.74	1277.85	0.00	416.57	0.00	1055.67	429.53	229.49	157.80	95.06	3281.80	20.69	1.11	0.00	0.00	0.80	0.00	0.00	3.31	9.33	74.50	0.00	368.54	223.88	0.00	0.00	223.56	106.08	2.83	0.00	28.10
2014	6.94	0.06	3.55	3040.63	0.00	446.92	14.74	1215.01	815.26	261.29	189.67	138.42	3807.03	21.25	1.50	23.82	12.52	0.71	297.74	16.65	5.63	8.49	95.85	1.34	382.59	265.75	5.50	11.02	449.17	112.42	2300.86	1.53	27.45
2015	7.55	0.00	2.78	1515.37	0.00	547.43	0.00	1219.26	806.57	329.61	209.24	129.68	4016.91	23.82	1.13	24.85	18.66	0.85	0.00	0.00	5.35	6.52	91.09	0.00	356.97	316.62	0.00	0.00	420.88	104.59	4.51	0.00	28.61
2016	7.34	0.34	4.34	3023.50	0.00	501.95	14.23	1077.72	859.32	277.81	177.58	131.25	4088.61	19.78	1.32	26.11	25.67	0.31	267.45	13.88	5.35	4.92	89.79	6.07	330.77	296.47	5.89	8.72	374.97	109.86	2304.82	1.96	23.89
2017	7.79	0.34	4.29	3733.14	0.59	584.24	14.21	1220.11	991.14	326.49	199.83	130.17	4800.77	22.14	1.45	29.23	29.95	0.71	319.10	14.74	15.97	6.88	89.70	6.28	374.28	331.43	8.79	11.39	395.35	114.03	2442.19	5.02	25.54

family	AUTOMOTIVE	BABY CARE	BEAUTY	BEVERAGES	BOOKS	BREAD/BAKERY	CELEBRATION	CLEANING	DAIRY	DELI	EGGS	FROZEN FOODS	GROCERY I	GROCERY II	HARDWARE	HOME AND KITCHEN I	HOME AND KITCHEN II	HOME APPLIANCES	HOME CARE	LADIESWEAR	LAWN AND GARDEN	LINGERIE	LIQUOR,WINE,BEER	MAGAZINES	MEATS	PERSONAL CARE	PET SUPPLIES	PLAYERS AND ELECTRONICS	POULTRY	PREPARED FOODS	PRODUCE	SCHOOL AND OFFICE SUPPLIES	SEAFOOD
year
2013	5.84	0.00	2.61	1260.96	0.00	418.76	0.00	1027.76	408.36	234.52	154.75	104.79	3308.63	21.95	1.01	0.00	0.00	0.75	0.00	0.00	3.23	9.39	71.10	0.00	377.75	224.65	0.00	0.00	221.59	96.68	5.35	0.00	25.37
2014	5.75	0.00	2.61	1461.64	0.00	416.34	0.00	1028.08	766.63	253.55	184.63	121.35	3614.11	21.44	1.02	0.00	0.00	0.56	0.00	0.00	4.98	8.94	85.83	0.00	361.61	266.44	0.00	0.00	405.27	103.47	5.66	0.00	23.63
2015	6.02	0.00	2.96	1915.16	0.00	551.60	0.00	1227.91	811.31	329.80	198.89	139.64	3989.90	26.46	1.06	26.11	20.73	0.65	127.72	0.00	5.47	8.20	95.21	1.69	365.24	314.39	0.06	0.00	415.49	102.44	5.68	0.00	27.05
2016	7.26	0.32	4.87	3369.27	0.00	551.90	14.25	1218.35	911.86	299.28	186.11	135.95	4898.71	22.04	1.36	27.10	28.28	0.45	315.92	14.96	5.47	6.24	82.89	6.77	363.50	363.47	6.84	10.67	399.37	108.67	2305.70	5.99	24.55
2017	7.63	0.23	5.30	3592.94	0.26	557.84	13.72	1308.01	970.85	323.21	202.54	138.61	4841.85	26.12	1.55	30.31	31.31	0.71	304.85	15.40	19.17	7.83	90.16	6.87	362.41	326.00	9.07	12.22	385.39	96.31	2404.91	11.71	23.17

2.2.1. Package¶

2.2.2. Nhập dữ liệu¶

2.2.3. Giao dịch¶

2.2.4. Giá dầu¶

2.3.1. Doanh thu¶

2.3.2. Dự báo bằng 0 (Zero Forecasting)¶

2.3.3. Kiểm tra các loại sản phẩm có hoạt động hay không¶

2.3.4. Sự kiện và dịp lễ¶

2.3.5. Các thuộc tính liên quan đến thời gian¶

2.3.6. Động đất có ảnh hưởng đến doanh thu của cửa hàng không?¶

Tháng 3¶

Tháng 4 - 5¶

Tháng 6¶

2.3.7. Hàm tự tương quan (ACF) và hàm tự tương quan riêng từng phần (PACF) cho họ sản phẩm¶

2.3.8. Đường trung bình trượt đơn giản¶

2.3.9. Đường trung bình trượt theo cấp số nhân¶

	date	store_nbr	family	sales
0	2017-08-16	1	BABY CARE	0
1	2017-08-17	1	BABY CARE	0
2	2017-08-18	1	BABY CARE	0
3	2017-08-19	1	BABY CARE	0
4	2017-08-20	1	BABY CARE	0
...	...	...	...	...
11	2017-08-27	54	LAWN AND GARDEN	0
12	2017-08-28	54	LAWN AND GARDEN	0
13	2017-08-29	54	LAWN AND GARDEN	0
14	2017-08-30	54	LAWN AND GARDEN	0
15	2017-08-31	54	LAWN AND GARDEN	0

	sales	SMA20_sales_lag16	SMA20_sales_lag30	SMA20_sales_lag60	SMA30_sales_lag16	SMA30_sales_lag30	SMA30_sales_lag60	SMA45_sales_lag16	SMA45_sales_lag30	SMA45_sales_lag60	SMA60_sales_lag16	SMA60_sales_lag30	SMA60_sales_lag60	SMA90_sales_lag16	SMA90_sales_lag30	SMA90_sales_lag60	SMA120_sales_lag16	SMA120_sales_lag30	SMA120_sales_lag60	SMA365_sales_lag16	SMA365_sales_lag30	SMA365_sales_lag60	SMA730_sales_lag16	SMA730_sales_lag30	SMA730_sales_lag60
sales	1.00	0.91	0.89	0.87	0.91	0.89	0.87	0.91	0.90	0.87	0.91	0.89	0.87	0.91	0.89	0.87	0.90	0.89	0.86	0.90	0.89	0.87	0.91	0.89	0.86
SMA20_sales_lag16	0.91	1.00	0.99	0.95	1.00	0.98	0.95	0.99	0.98	0.95	0.99	0.98	0.95	0.99	0.98	0.95	0.98	0.97	0.95	0.97	0.97	0.95	0.98	0.98	0.95
SMA20_sales_lag30	0.89	0.99	1.00	0.96	0.99	1.00	0.96	1.00	0.99	0.97	0.99	0.99	0.97	0.99	0.99	0.96	0.99	0.98	0.96	0.97	0.97	0.96	0.98	0.98	0.96
SMA20_sales_lag60	0.87	0.95	0.96	1.00	0.96	0.98	1.00	0.98	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.98	0.98	0.97	0.97	0.98	0.98	0.98
SMA30_sales_lag16	0.91	1.00	0.99	0.96	1.00	0.99	0.96	1.00	0.99	0.96	0.99	0.99	0.96	0.99	0.98	0.96	0.99	0.98	0.96	0.98	0.97	0.96	0.98	0.98	0.96
SMA30_sales_lag30	0.89	0.98	1.00	0.98	0.99	1.00	0.98	1.00	1.00	0.98	1.00	0.99	0.98	0.99	0.99	0.97	0.99	0.99	0.97	0.98	0.98	0.97	0.98	0.98	0.98
SMA30_sales_lag60	0.87	0.95	0.96	1.00	0.96	0.98	1.00	0.98	0.99	1.00	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.98	0.98	0.98	0.98	0.98	0.98
SMA45_sales_lag16	0.91	0.99	1.00	0.98	1.00	1.00	0.98	1.00	1.00	0.98	1.00	0.99	0.98	0.99	0.99	0.98	0.99	0.99	0.97	0.98	0.98	0.97	0.98	0.98	0.98
SMA45_sales_lag30	0.90	0.98	0.99	0.99	0.99	1.00	0.99	1.00	1.00	0.99	1.00	1.00	0.99	1.00	0.99	0.98	0.99	0.99	0.98	0.98	0.98	0.98	0.98	0.98	0.98
SMA45_sales_lag60	0.87	0.95	0.97	0.99	0.96	0.98	1.00	0.98	0.99	1.00	0.99	0.99	1.00	0.99	1.00	0.99	0.99	0.99	0.99	0.98	0.98	0.98	0.98	0.98	0.98
SMA60_sales_lag16	0.91	0.99	0.99	0.99	0.99	1.00	0.99	1.00	1.00	0.99	1.00	1.00	0.99	1.00	0.99	0.98	0.99	0.99	0.98	0.98	0.98	0.98	0.98	0.98	0.98
SMA60_sales_lag30	0.89	0.98	0.99	0.99	0.99	0.99	0.99	0.99	1.00	0.99	1.00	1.00	0.99	1.00	1.00	0.99	1.00	0.99	0.98	0.98	0.98	0.98	0.98	0.98	0.98
SMA60_sales_lag60	0.87	0.95	0.97	0.99	0.96	0.98	0.99	0.98	0.99	1.00	0.99	0.99	1.00	0.99	1.00	1.00	1.00	1.00	0.99	0.99	0.99	0.98	0.98	0.98	0.98
SMA90_sales_lag16	0.91	0.99	0.99	0.99	0.99	0.99	0.99	0.99	1.00	0.99	1.00	1.00	0.99	1.00	1.00	0.99	1.00	1.00	0.99	0.99	0.99	0.98	0.99	0.98	0.98
SMA90_sales_lag30	0.89	0.98	0.99	0.99	0.98	0.99	0.99	0.99	0.99	1.00	0.99	1.00	1.00	1.00	1.00	0.99	1.00	1.00	0.99	0.99	0.99	0.98	0.98	0.99	0.98
SMA90_sales_lag60	0.87	0.95	0.96	0.99	0.96	0.97	0.99	0.98	0.98	0.99	0.98	0.99	1.00	0.99	0.99	1.00	1.00	1.00	1.00	0.99	0.99	0.99	0.98	0.98	0.99
SMA120_sales_lag16	0.90	0.98	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00	1.00	0.99	0.99	0.99	0.99	0.99	0.99	0.99
SMA120_sales_lag30	0.89	0.97	0.98	0.99	0.98	0.99	0.99	0.99	0.99	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00	1.00	0.99	0.99	0.99	0.99	0.99	0.99
SMA120_sales_lag60	0.86	0.95	0.96	0.98	0.96	0.97	0.99	0.97	0.98	0.99	0.98	0.98	0.99	0.99	0.99	1.00	0.99	1.00	1.00	0.99	0.99	0.99	0.99	0.99	0.99
SMA365_sales_lag16	0.90	0.97	0.97	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.99	0.99	0.99	0.99	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00
SMA365_sales_lag30	0.89	0.97	0.97	0.97	0.97	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.99	0.99	0.99	0.99	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00
SMA365_sales_lag60	0.87	0.95	0.96	0.97	0.96	0.97	0.98	0.97	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.99	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00
SMA730_sales_lag16	0.91	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.99	0.98	0.98	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00
SMA730_sales_lag30	0.89	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.99	0.98	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00
SMA730_sales_lag60	0.86	0.95	0.96	0.98	0.96	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.98	0.99	0.99	0.99	0.99	1.00	1.00	1.00	1.00	1.00	1.00

family	AUTOMOTIVE	BABY CARE	BEAUTY	BEVERAGES	BOOKS	BREAD/BAKERY	CELEBRATION	CLEANING	DAIRY	DELI	EGGS	FROZEN FOODS	GROCERY I	GROCERY II	HARDWARE	HOME AND KITCHEN I	HOME AND KITCHEN II	HOME APPLIANCES	HOME CARE	LADIESWEAR	LAWN AND GARDEN	LINGERIE	LIQUOR,WINE,BEER	MAGAZINES	MEATS	PERSONAL CARE	PET SUPPLIES	PLAYERS AND ELECTRONICS	POULTRY	PREPARED FOODS	PRODUCE	SCHOOL AND OFFICE SUPPLIES	SEAFOOD
year
2013	5.89	0.00	2.72	1297.24	0.00	431.05	0.00	1052.44	419.08	238.23	156.45	120.79	3489.13	21.83	0.95	0.00	0.00	0.71	0.00	0.00	3.43	8.85	75.67	0.00	389.95	216.12	0.00	0.00	232.16	100.72	6.12	0.00	20.99
2014	5.72	0.00	2.46	1458.94	0.00	450.82	0.00	1029.45	757.82	260.01	182.69	122.18	3655.15	22.99	1.03	0.00	0.00	0.41	0.00	0.00	4.35	8.60	94.37	0.00	383.37	270.10	0.00	0.00	409.03	101.04	5.75	0.00	23.97
2015	6.52	0.18	3.79	3178.15	0.00	549.95	14.74	1177.30	829.06	328.15	196.27	139.39	3903.51	25.12	1.30	26.92	31.28	0.12	288.21	16.89	5.27	7.83	101.28	1.53	376.68	297.36	5.87	8.72	411.02	108.23	2346.14	1.33	25.85
2016	6.25	0.35	4.55	2806.97	0.00	550.81	13.96	1108.43	863.92	280.66	182.98	129.32	4170.43	21.84	1.08	28.62	25.08	0.29	260.83	14.07	5.66	4.43	85.12	5.67	350.66	292.74	5.95	8.50	373.23	104.96	2239.03	1.55	23.18
2017	7.38	0.23	5.43	3580.13	0.08	555.64	13.80	1228.70	932.53	317.70	202.11	137.49	4702.43	29.26	1.40	33.93	30.85	0.64	286.55	14.65	17.85	6.78	95.40	6.19	388.38	319.44	9.22	11.34	406.58	90.90	2403.45	1.58	19.36